unable to disable -r in run_tests.py

[doko42]

BPO	17774
Nosy	@doko42, @csabella

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = None
created_at = <Date 2013-04-17.13:03:01.616>
labels = ['3.8', 'type-bug', 'tests']
title = 'unable to disable -r in run_tests.py'
updated_at = <Date 2019-05-18.01:19:59.944>
user = 'https://github.com/doko42'

bugs.python.org fields:

activity = <Date 2019-05-18.01:19:59.944>
actor = 'cheryl.sabella'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Tests']
creation = <Date 2013-04-17.13:03:01.616>
creator = 'doko'
dependencies = []
files = []
hgrepos = []
issue_num = 17774
keywords = []
message_count = 2.0
messages = ['187160', '342781']
nosy_count = 2.0
nosy_names = ['doko', 'cheryl.sabella']
pr_nums = []
priority = 'normal'
resolution = None
stage = 'needs patch'
status = 'open'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue17774'
versions = ['Python 3.8']

[doko42]

it is not possible to disable -r in run_tests.py. Other options like -u or -j can be overwritten, but not removing -r.

[csabella]

Since -r doesn't usually take a value, what would be the best way to implement this? The current implementation always adds -r, so not including '-r' would need to retain the current behavior of randomization. Would something like '--no-random' be a good way to implement this? That would be different than how the program handles the -u or -j because the '--no-random' would have to be removed from regrtest_args.

[iritkatriel]

run_tests.py no longer exists.

[terryjreedy]

The default today is to run tests sequentially in alphabetical order. Passing -r randomizes the sequential order. Passing -jn runs tests in what appears to be a deterministic but undocumented and likely subject-to-change non-alphabetical order. Passing both randomizes a parallel run. So it seems the request above is fulfilled and is appropriately closed. It does not matter what the situation was in 3.8. Starting parallel tests alphabetically would not result in them finishing and being reported alphabetically.